Our goal is to survey and achieve a systems level understanding of the state space and circuitry of macrophage activation at the cell population and single-cell levels. Using the resulting dense phenotypic data sets, including transcriptomes, chromatin states, and signaling statuses, we aim to develop and apply computational approaches to disentangle biological information from measurement noise, as well as to infer state-specific gene regulatory networks. Thus far we have been using human blood monocyte derived macrophages (MDM) as pimary models; other related models, including THP1 (a human monocyte cell line), RAW264.7 (a mouse macrophage cell line), and mouse bone marrow derived macrophages (BMDM) have also been used. These systems, particularly the RAW and BMDMs, are also being investigated by colleagues at the Laboratory of Systems Biology through which we can achieve synergy by integrating information from additional levels (e.g., protein-protein interactions, protein expression), time scales (e.g., connecting signaling to transcriptional responses), and perturbations (RNAi). The following are highlights of our current research activities in this area: 1. Apply experimental genomics approaches to measure responses before and after perturbations and use (and develop new ones if necessary) computational approaches to process and integrate such data to systematically infer gene regulatory networks. We are using a combination of RNAseq, microarray and Fluidigm the former can provide detailed transcriptomic information including the abundance of non-coding genes, alternative splicing isoforms and rare transcripts; microarrays can be applied to a larger number of conditions/samples because they are less expensive; Fluidigm offers inexpensive profiling of a large number of samples by focusing on a carefully selected panel of transcripts. Some example efforts include: a. In collaboration with other labs at the LSB, we have assayed the dose response of BMDMs to TLR ligands as measured by phosphorylation of key signaling factors and the transcriptome. We observed that NF-kB and MAPK pathways are activated at quantitatively distinct thresholds, thus suggesting that inflammatory responses and steady-state microbe sensing are uncoupled (see Gottschalk et al. 2016). b. In collaboration with the Germain and Fraser labs, we have assayed the transcriptomic (mRNA) and microRNA responses of BMDMs in response to dose combinations of TLR ligands. We observed largely non-linear responses across the dose range when pairs of ligands are applied, suggestive of synergistic or antagonistic interactions among the components of the signaling and transcriptional networks. We are particularly interested in integrating the transcriptional data with signaling status to infer the connection between signaling events to downstream transcriptional regulation. c. In collaboration with Iain Frasers lab, we used RNAseq to assay macrophages before and after siRNA knockdown of promising hits from genome-wide siRNA screening. In particular, we are focusing on factors that regulate mRNA splicing. d. In another project we stimulated RAW cells with LPS (a prototypical TLR activator) and generated RNAseq data with deep coverage (160 million reads per sample). By utilizing the deep coverage provided by RNAseq, we have developed computational approaches to characterize the extent of non-conserved unexpected splicing events (USE), many of which have postulated to be results of noisy splicing. We found that USEs are prevalent across macrophages in both resting and stimulated conditions. The extent of USE is highly variable across genes, with certain pathway exhibiting significantly different levels of USEs. e. We have generated RNAseq data on both coding and non-coding RNAs, enhancer statuses (as indicated by h3k27ac levels), and chromatin accessibility (via ATACseq) across a number of human macrophage activation conditions involving stimulation by individual or combination of cytokines. These data are being used to understand the enhancer, chromatin state, transcriptomic landscape and splicing repertoire of macrophage activation states. These human data sets are also being used to assess the functional implications of disease-associated genetic variants e.g., asking whether disease-associated variants might directly impact transcription factor binding. 2. Use single-cell gene expression assays to measure the transcriptome in individual cells before and after cellular activation. We combined flow cytometry with single-cell based PCR (Fluidigm) to assay the expression of 100 genes/proteins. Using human blood derived macrophage as a model, we have generated data from hundreds of single cells across several stimulation conditions. We are particularly interested in assaying expression heterogeneity under different conditions because we would like to compare environmentally induced phenotypic differences at both the cell population and individual cell levels. We are also collaborating with other colleagues by utilizing our single cell profiling approach to help answer questions about cellular heterogeneity, for example, in the context innate lymphoid cell development (with the lab of Jeff Zhu of the Laboratory of Immunology.) 3. To help assess cell-to-cell variations in a robust manner, we have incorporated measuring random samples of 10 cells (aka stochastic profiling) on top of single-cell data and developed a novel Bayesian computational inference approach to integrate both data types for quantifying cell-to-cell variations. Our approach can also be used on single- or k-cell data alone. We have evaluated the use of different data types (i.e., single- and k-cell separately, or combining both) and provided scenarios under which individual data types give robust inference of cellular heterogeneity within single conditions as well as analysis of differences across two conditions. We also provide an R package to facilitate the application of our approach by others (see Narayanan et al. 2016). 4. Utilize the natural variation across single cells to construct condition-specific gene regulatory networks. Using both single- and 10-cell data described above, we have examined state-specific gene-gene correlation networks in human monocyte-derived macrophages. We observed significant differences in gene-gene correlations across activation conditions. In particular, activating transcription factor 2 (ATF2), a component of the activating protein (AP)-1 complex, is uniquely correlated to a set of genes containing both classically inflammatory and anti-inflammatory genes after interleukin (IL)-10 activation. ATF2/AP1 is primarily thought of as a factor activated downstream of inflammatory (e.g. LPS) activation, its role in IL-10 stimulated conditions has not been described. Microscopy and flow cytometry experiments revealed that a higher level of phosphorylated nuclear ATF2 is present after IL-10 activation. ChIP-Seq analysis indicated higher levels of h3k27ac marks in ATF2-motif bearing enhancers nearby the genes correlated with ATF2-across single cells. Anti-ATF2 ChIP experiments further suggested direct ATF2 binding within these regions. In addition to uncovering a novel circuit involving ATF2 and its unique partners in response to IL-10 activation in human macrophages (despite IL-10 being classically thought of as an anti-inflammatory activator and ATF2 as downstream of inflammatory and stress signals), our analyses provide a conceptual framework and a concrete example on the signaling and transcriptional network underpinnings of how cell-to-cell variations are propagated and how propagation behavior can differ from environment to environment.